Loading [MathJax]/extensions/MathMenu.js
Strategies for Fast I/O Throughput in Large-Scale Climate Modeling Applications | IEEE Conference Publication | IEEE Xplore

Strategies for Fast I/O Throughput in Large-Scale Climate Modeling Applications


Abstract:

Large-scale HPC applications are highly data-intensive with significant times spent in I/O operations. Many large-scale scientific applications do not adequately optimize...Show More

Abstract:

Large-scale HPC applications are highly data-intensive with significant times spent in I/O operations. Many large-scale scientific applications do not adequately optimize the I/O operations, leading to overall poor performance. In this work, we have developed two main strategies for providing fast I/O throughput for an important climate modeling application, namely, Regional Ocean Modeling System (ROMS) that uses NetCDF for I/O operations. The strategies include load balancing the I/O operations and selective writing of data. We have also implemented file striping to improve I/O performance. Our experiments with up to 1440 processor cores and 5 days of simulations showed that our load balancing strategy resulted in about 27 % decrease in execution times over the default executions, our selective writing strategy resulted in a further decrease of about 30 % and the optimized file striping resulted in a further decrease of about 12 % in execution times. All the strategies combined together improved the overall performance of the application by about 70 %.
Date of Conference: 18-21 December 2023
Date Added to IEEE Xplore: 05 April 2024
ISBN Information:

ISSN Information:

Conference Location: Goa, India
References is not available for this document.

I. Introduction

ROMS [1] is an ocean model widely used by the scientific community for a wide range of climate applications. It can work standalone as well as it can be coupled to atmospheric and/or wave models. It is built based on the Earth System Modeling Framework (ESMF) [2] which provides high per-formance and flexibility for coupling climate and related other scientific applications. The format of the input and output data of the model are Network Common Data Form (NetCDF) [3] which helps to interchange the data in a user-friendly way.

Select All
1.
Wikiroms, [online] Available: https://www.myroms.org/wiki/Documentation\_Portal.
2.
Esmf, [online] Available: https://www.earthsystemcog.org/projects/esmf/.
3.
Netcdf, [online] Available: https://www.unidata.ucar.edu/software/netcdf/.
4.
Z. e. a. Liu, "Profiling and Improving I/O Performance of a Large-scale Climate Scientific Application", 2013 22nd International Conference on Computer Communication and Networks (ICCCN), pp. 1-7, 2013.
5.
Geos, [online] Available: https://gmao.gsfc.nasa.gov/systems/geos5/.
6.
J. Lofstead, S. Klasky, K. Schwan, N. Podhorszki and C. Jin, "Flexible I/O and Integration for Scientific Codes through the Adaptable I/O System (ADIOS)", Proceedings of the 6th international workshop on Challenges of large applications in distributed environments, pp. 15-24, 2008.
7.
H. Tang, X. Zou, J. Jenkins, D. A. Boyuka, S. Ranshous, D. Kimpe, et al., "Improving Read Performance with On-line Access Pattern Analysis and Prefetching", European Conference on Parallel Processing, pp. 246-257, 2014.
8.
R. K. Rew, B. Ucar and E. Hartnett, "Merging NetCDF and HDF5", 20th Int. Conf. on Interactive Information and Processing Systems, 2004.
9.
M. J. Folk, R. Rew, M. Yang, E. Hartnett, R. E. McGrath and Q. Koziol, "NetCDF-4: Combining NetCDF and HDF5 Data", AGU Fall Meeting Abstracts, 2003.
10.
W. W. Dai, A. J. Scannapieco, F. L. Cochran, C. Chang, P. M. Weber, E. L. Sandford, et al., "Buffering I/O for Data Management in Multi-Physics Simulations", Proceedings of the High Performance Computing Symposium, pp. 1-8, 2013.
11.
B. Behzad, J. Huchette, H. Luu, R. Aydt, Q. Koziol, M. Prabhat, et al., "Auto-tuning of Parallel I/O Parameters for HDF5 Applications", 2012 SC Companion: High Performance Computing Networking Storage and Analysis, pp. 1430-1430, 2012.
12.
K. Gao, W.-k. Liao, A. Choudhary, R. Ross and R. Latham, "Combining I/O Operations for Multiple Array Variables in Parallel NetCDF", 2009 IEEE International Conference on Cluster Computing and Workshops, pp. 1-10, 2009.
13.
K. Gao, W.-k. Liao, A. Nisar, A. Choudhary, R. Ross and R. Latham, "Using Subfiling to Improve Programming Flexibility and Performance of Parallel Shared-file I/O", 2009 International Conference on Parallel Processing, pp. 470-477, 2009.
14.
Y. Tsujita, A. Hori, T. Kameyama, A. Uno, F. Shoji and Y. Ishikawa, "Improving Collective MPI-IO Using Topology-aware Stepwise Data Aggregation with I/O Throttling", Proceedings of the International Conference on High Performance Computing in Asia-Pacific Region, pp. 12-23, 2018.
15.
R. Thakur, W. Gropp and E. Lusk, "Data Sieving and Collective I/O in ROMIO", Proceedings. Frontiers' 99. Seventh Symposium on the Frontiers of Massively Parallel Computation, pp. 182-189, 1999.
16.
S. Snyder, P. Carns, K. Harms, R. Ross, G. K. Lockwood and N. J. Wright, "Modular HPC I/O Characterization with Darshan", 2016 5th Workshop on Extreme-Scale Programming Tools (ESPT), pp. 9-17, 2016.
17.
B. Xie, J. Chase, D. Dillow, O. Drokin, S. Klasky, S. Oral, et al., "Characterizing Output Bottlenecks in a Supercomputer", SC'12: Proceedings of the International Conference on High Performance Computing Networking Storage and Analysis, pp. 1-11, 2012.
18.
P. Carns, K. Harms, W. Allcock, C. Bacon, S. Lang, R. Latham, et al., "Understanding and Improving Computational Science Storage Access through Continuous Characterization", ACM Transactions on Storage (TOS), vol. 7, no. 3, pp. 1-26, 2011.
19.
Netcdf, [online] Available: http://meteora.ucsd.edu/rvpierce/ncview_home_page.html.
20.
Hdf5, [online] Available: https://www.hdfgroup.org/.
21.
S. M. Pargaonkar and P. N. Vinayachandran, "Wind Forcing of the Ganga-Brahmaputra River Plume", Ocean Dynamics, vol. 71, no. 2, pp. 125-156, Feb. 2021.
22.
Craypat, [online] Available: https://docs.nersc.gov/programming/performance-debugging-tools/craypat/.
23.
Lustre, [online] Available: https://www.nics.tennessee.edu/computing-resources/file-systems/io-Iustre-tips.

References

References is not available for this document.